Phizenix-posted 1 day ago
$70 - $75/Yr
Full-time • Mid Level
Sunnyvale, CA

We are seeking a modeling Engineer to develop high-level models of complex SoC hardware. The virtual platforms combine models of custom hardware accelerators for vision, 2D and 3D graphics, machine learning and more, within a multi-core, multi-level memory hierarchy SoC architecture, and serve as the primary simulation vehicle for system software and firmware. The ideal candidate will be proficient in hardware simulation using C++, and understand the firmware development processes. Responsibilities Design and develop SystemC TLM models to accurately represent the SoC architecture integrating emulated processors, DSPs, Network-on-Chip, DMA and memory controllers, etc… Integrate first-party and vendor models into the Virtual Platform, develop automated workflows to ensure register-level accuracy and complete connectivity at the SoC level, minimizing manual intervention and enabling continuous integration. Collaborate with silicon architects, digital designers and verification engineers to design and develop high-fidelity, fast C++ models for first-party IP. Coordinate virtual platforms with hardware development programs, validating multiple SoCs and architectural changes with system software and firmware engineering, enabling end-to-end silicon validation test frameworks. Enhance the virtual platforms to enable SoC and system architecture exploration by instrumenting models for power and performance metrics, allowing for data-driven design decisions and trade-off analysis to optimize system performance and power consumption.

  • Design and develop SystemC TLM models to accurately represent the SoC architecture integrating emulated processors, DSPs, Network-on-Chip, DMA and memory controllers, etc…
  • Integrate first-party and vendor models into the Virtual Platform, develop automated workflows to ensure register-level accuracy and complete connectivity at the SoC level, minimizing manual intervention and enabling continuous integration.
  • Collaborate with silicon architects, digital designers and verification engineers to design and develop high-fidelity, fast C++ models for first-party IP.
  • Coordinate virtual platforms with hardware development programs, validating multiple SoCs and architectural changes with system software and firmware engineering, enabling end-to-end silicon validation test frameworks.
  • Enhance the virtual platforms to enable SoC and system architecture exploration by instrumenting models for power and performance metrics, allowing for data-driven design decisions and trade-off analysis to optimize system performance and power consumption.
  • B.S. degree in Computer Science or Electrical Engineering or equivalent experience.
  • 2+ years experience in hardware model simulation, virtual platform, performance modeling of complex SoCs or high-fidelity hardware accelerators.
  • High proficiency in modern C++ in the domains of chip-design, electronic design automation or simulation.
  • General familiarity with SoC components: embedded processors such as ARM A/M series, Risc-V, DSP, DMA, Cache Hierarchy, DRAM, Network-on-chip, AMBA protocols. Extensive experience in at least one of these areas.
  • Experience with modern buildframeworks and continuous integration systems, such as CMake, Bazel and CI frameworks such as Jenkins, GitLab CI/CD.
  • Experience with debugging and profiling tools, such as GDB or other debuggers
  • Experience with the SystemC/TLM library
  • Experience with virtual platform development tools and frameworks, such as Synopsys Virtualizer, Cadence Virtual Platform, Imperas OVP, or ARM Fast Models
  • Familiarity with processor/DSP architectures, such as ARM, RISC-V, and XTensa
  • Familiarity with NoC, MMU, address translations, and cache modeling
  • Familiarity with the standard C++ concurrency support library: threads, atomic operations, memory ordering, etc…
  • Proficiency in Python to automate design flows, creation of collateral data
  • Experience with high level C/C++ synthesis (HLS)
  • Working knowledge of Verilog
© 2024 Teal Labs, Inc
Privacy PolicyTerms of Service